Mike Robinson knows how to make an exit.
Well, he knows how long it takes to make an exit.
Robinson, an associate research professor at Old Dominion University's Virginia Modeling, Analysis and Simulation Center (VMASC), works on evacuation models for emergencies that force community evacuations, including natural disasters and man-made events such as terrorist attacks.
"I wouldn't say I'm always scanning for an exit, but I look to see, 'Could people get out of here?'" he said.
ODU is a leader in the disaster response field.
When the Camp Fire, the deadliest and most destructive wildfire in California's history, devastated communities in 2018, Robinson was part of a team of evacuation experts hired by Pacific Gas and Electric Company (PG&E) to assess wildfire evacuation plans in the state. A faulty power line owned by PG&E is believed to have caused the fire which nearly destroyed the towns of Paradise and Concow and killed 85 people.
"When we got out there, we were looking at these communities and so many of them did not have what we would believe to be acceptable evacuation plans," Robinson said. "Part of that is because modeling wildfires is very complex, so evacuation studies had not been emphasized. Smaller communities were especially affected because of the cost to run simulations."
So, Robinson and the modeling team at VMASC set out to create a tool for these vulnerable communities.
The Fast Local Emergency Evacuation Times (FLEET) model uses precise census data and smaller road networks - like the ones in these towns - to estimate the time it takes to evacuate a given area.
On a page that looks sort of like Google Maps, users select the area they want to leave, and within minutes, FLEET delivers an estimated evacuation time.
"FLEET filled a void, and I'm happy to say that people are using it," Robinson said.
Released to the public in August 2021, the program has been used by communities in 24 states.
FLEET builds off the Real Time Evacuation Planning Model (RtePM) which VMASC developed with the Applied Physics Laboratory at Johns Hopkins University in 2012.
In November 2021, Robinson was once again contacted for his knowledge on evacuation modeling by a group called Westside Watch in Colorado Springs, Colorado. They asked him to run the FLEET model for several areas of the city in hopes of motivating Colorado Springs' city council to do a larger evacuation study.
"Mike's expertise in evacuation modeling has been critical to educating our community and leadership on why this process is so needed, especially in the new normal of dire fires we are experiencing across the country," said Dana Duggan, board member for Westside Watch. "We may not be able to stop the fires, but we sure can use cutting-edge tools like FLEET to ensure that people can get out and improve our ingress and egress in advance of these fires."
Not long after Robinson and the team presented to city council, the Marshall Fire destroyed hundreds of homes near Boulder, Colorado. In late March, another wildfire near Boulder forced 19,000 people to evacuate their homes.
"People say, 'It's not going to happen to me,' and they don't prepare for emergencies," Robinson said. "If there is a positive takeaway from people losing their homes in such a terrible way, perhaps it's that others will wake up and say, 'We need to plan for this.'"
Although Robinson and other researchers at VMASC aren't making those decisions, they are dedicated to developing the tools that communities and leaders can use to plan for emergencies.
"We just want people to be safe," he said.
VMASC is a multidisciplinary applied research and enterprise research facility of Old Dominion University, located in the Tri-Cities Center in Suffolk. Staffed by more than two dozen research faculty and project scientists, they provide modeling and simulation, analytic research, and technological support for partners across various industry, government, and community sectors: including, health care, cybersecurity, strategic defense, transportation and infrastructure, usability, and instructional design.